DBI-616 Build Question for Perl 5.8.2 on AIX 5.3

2011-12-03 Thread Curtis Leach
Guys,

 

I'm building the DBI on AIX and I'm seeing something I'm not sure I like
in the test results.  I'm using the pre-built perl that comes with AIX
5.3 and IBM's C compiler that was used to build it.

 

To cut down on most of the miscellaneous messages about skipping
SQL-Statement, I went and installed SQL-Statement-1.33 before I
attempted to build DBI again.  But even with cleaner log files, it did
nothing for the log entries I was worried about.

 

I saved & reviewed the output of "perl Makefile.PL" & "make", but didn't
see anything in there that looked suspicious.  (logs available on
request)

 

Here's a sample of the messages I'm talking about.

t/35thrcloneAttempt to free unreferenced scalar: SV
0x3081c914 during global destruction.

Attempt to free unreferenced scalar: SV 0x306b9244 during global
destruction.

Attempt to free unreferenced scalar: SV 0x305501c4 during global
destruction.

Attempt to free unreferenced scalar: SV 0x3037bcf4 during global
destruction.

Ok

 

These messages appear for several test cases.  But the log ends with a
message that implies all is OK:

t/zvxnp_85gofer.ok

18/234 skipped: caching of metadata methods returning sth not
yet implemented

All tests successful, 38 tests and 553 subtests skipped.

Files=178, Tests=7937, 147 wallclock secs (111.78 cusr + 25.73 csys =
137.51 CPU)

PERL_DL_NONLAZY=1 /usr/bin/perl "-Iblib/lib" "-Iblib/arch"
test.pl

test.pl

DBI test application $Revision: 12537 $

Switch: DBI 1.616 by Tim Bunce, 1.616

Available Drivers: DBM, ExampleP, File, Gofer, Proxy, Sponge

dbi:ExampleP:: testing 3 sets of 20 connections:

Connecting... 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Disconnecting...

Connecting... 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Disconnecting...

Connecting... 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20

Disconnecting...

connect 20 and disconnect them, 3 times: 0.0050s / 60 = 0.0001s

Testing handle creation speed...

2 NullP sth/s perl 5.008002 aix-thread-multi (cc_r 5.0.2.0 -O)
0.36s

 

test.pl done

Target "test" is up to date.

 

Are these messages about attempts to free unreferenced scalars
important?  It sounds like I'm hitting some type of problem that's being
ignored in the pass/fail tests.

 

If I can ignore it, I'm good to go, but I'm a bit nervous about this
warning.

 

I can provide the output of "perl Makefile.PL", "make", and "make test"
on request.

 

Curtis

 



Maintaining simultaneous support for two Oracle versions in DBI

2011-12-03 Thread Scott Smith
My workplace is transitioning from Oracle version 9 to version 11. I 
would like to build the DBD driver to support connecting to the Oracle 
11 database. However, until all the data are migrated from the old 
database to the new one, I would need to maintain connectivity to the 
old Oracle 9 database. Since the driver and shared objects built for the 
new database would be incompatible, I would need to build separate 
objects for the Oracle 11 connection.


Some of the items I can identify are:

The module should be Oracle11.pm rather than Oracle.pm (called as 
DBI->connect("dbi:Oracle11:", ...) rather than 
DBI->connect("dbi:Oracle:", ...)).
The shared object and bootstrap files as Oracle11.so and Oracle11.bs 
respectively. (Should these go in the same directory or should they be 
in the directory Oracle11 and does the package name in Oracle11.pm have 
to be changed to DBD::Oracle11 in order to find things there?)


I have some idea of what I need to change in Makefile.PL in order to 
achieve at least part of this. What I have so far is changing references 
to Oracle.pm to Oracle11.pm and $opts{NAME} from DBD::Oracle to Oracle 
11. Am I right about the second? Are there others I should have included?


Thanks for any help,
Scott Smith